Memory MCP Server

Overview

This project implements a simple Memory MCP (Memory Control Protocol) server using fastmcp. The server provides a set of tools for storing, retrieving, and managing key‑value pairs in a JSON file. It also supports namespaces and wildcard key listing.

Installation

# Create a virtual environment
python -m venv .venv
source .venv/bin/activate  # On Windows use .venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

Running the Server

python memory_server.py

The server runs with a stdio transport and will start listening for tool calls.

Running the Client Example

In a separate terminal (with the same virtual environment activated) run:

python memory_client.py

The client will:

1. Save a key/value pair.
2. Retrieve it.
3. List all keys.
4. Delete the key.
5. Demonstrate namespace usage.

Output Example

save result: True
get result: {'key': 'foo', 'value': {'bar': 1}, 'timestamp': 1701234567.890}
list_keys result: ['foo']
delete result: True
get_by_namespace result: [{'key': 'ns1:baz', 'value': 'qux', 'timestamp': 1701234568.123}]

Persistence

All data is stored in memory_data.json in the same directory as the server. Each entry contains the key, value, and a Unix timestamp.

Dependencies

• fastmcp
• pydantic
• python-dotenv

All dependencies are specified in requirements.txt.